Create Captioned Podcast Audiograms in n8n (Zvid + ElevenLabs + OpenRouter)

Create captioned vertical podcast audiograms in n8n (Zvid + ElevenLabs + OpenRouter)

This n8n workflow automatically pulls your newest RSS episode, transcribes it with ElevenLabs for accurate word-level timestamps, selects highlight moments via OpenRouter, and renders branded vertical captioned audiogram MP4s using Zvid.

What this workflow does

• Runs on demand or on a schedule to select the newest audio episode from your RSS feed.
• Downloads a bounded audio sample and verifies its actual size before transcription (default sample size: 4 MiB; configurable via maxAudioBytes).
• Transcribes with ElevenLabs Scribe, using its word timestamps to produce accurate captions and clip boundaries.
• Uploads the same audio to Zvid and uses OpenRouter to pick moments based on transcript word positions.
• Builds, validates, and quotes each branded audiogram, then returns completed rendered outputs by default.
• Returns editor drafts with dryRun=true (AI/transcription charges may still apply), along with quotes.
• Renders sequentially by default; retries explicit capacity rejections within a deadline.

Use cases

• Podcast teams that want consistent, fast social-ready vertical audiogram clips for every new RSS episode.
• SaaS operators producing episode promos who need captioned, branded renders without manual editing.
• n8n automation engineers who want a production workflow combining transcription, highlight selection, and video rendering.

Technical details

• n8n nodes: if, set, code, wait, @zvid/n8n-nodes-zvid (Zvid), sticky note
• Zvid credential: Base URL https://api.zvid.io and API key from https://app.zvid.io/api-keys
• ElevenLabs integration: Header Auth “xi-api-key” with Speech to Text permission (Transcribe → Scribe)
• OpenRouter integration: configured to pick moments using transcript word positions (feedUrl, showNameOverride, handle/branding in Config)